Abstract
If a thin luminance contour and a larger solid disk are dichoptically flashed to opposite eyes, a dark hole is seen in the disk. The contour masks the disk's interior by blocking brightness information which normally “fills-in” from disk's edge (Paradiso and Nakayama, 1991). Using color versions of this paradigm, we find that a color contour mask can weaken if it is part of a larger global configuration. Here, adding to a mask decreases its blocking power, perhaps because the addition provides cues that the configuration is a separate, occluding surface. Method: A solid red disk (target) and thin green contour (mask) are flashed simultaneously to opposite eyes. The mask is either square or #-shaped (the same square with arms extending outside the target). The stimulus repeats until the subject responds. The task is to adjust the intensity of a peripheral color patch to match the redness perceived within the center of the mask. Result: subjects found the # to mask more weakly than the square, reporting the central redness to be significantly more intense in # trials. Some subjects reported that the # configuration often failed to mask the center at all, while the square configuration was a consistently strong mask. When the experiment was run using achromatic stimuli, subjects saw no significant difference between masks, some found the # to be the stronger mask. Discussion: Previously, we showed that the slow color filling-in underlying Troxler fading can jump over luminance contours and fill discretely to remote areas. The ability of a contour to block filling-in is based not only on its local contrast properties, but also on global surface segregation. Here, we extend this finding to the rapid filling-in underlying normal perception. The #-shaped mask is seen to be separate from the disk's surface, which weakens the ability of the square portion of the # to block the filling-in of the disk. However, this does not seem to apply to the filling-in of achromatic brightness.